句法不可比平行句的过滤器

Q2 Arts and Humanities

Linguistics in the Netherlands Pub Date : 2019-11-05 DOI:10.1075/avt.00029.kro

Martin Kroon, S. Barbiers, J. Odijk, S. V. D. Pas

{"title":"句法不可比平行句的过滤器","authors":"Martin Kroon, S. Barbiers, J. Odijk, S. V. D. Pas","doi":"10.1075/avt.00029.kro","DOIUrl":null,"url":null,"abstract":"\n Massive automatic comparison of languages in parallel corpora\n will greatly speed up and enhance comparative syntactic research. Automatically\n extracting and mining syntactic differences from parallel corpora requires a\n pre-processing step that filters out sentence pairs that cannot be compared\n syntactically, for example because they involve “free” translations. In this\n paper we explore four possible filters: the Damerau-Levenshtein distance between\n POS-tags, the sentence-length ratio, the graph-edit distance between dependency\n parses, and a combination of the three in a logistic regression model. Results\n suggest that the dependency-parse filter is the most stable throughout language\n pairs, while the combination filter achieves the best results.","PeriodicalId":35138,"journal":{"name":"Linguistics in the Netherlands","volume":" ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2019-11-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"1","resultStr":"{\"title\":\"A filter for syntactically incomparable parallel\\n sentences\",\"authors\":\"Martin Kroon, S. Barbiers, J. Odijk, S. V. D. Pas\",\"doi\":\"10.1075/avt.00029.kro\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\n Massive automatic comparison of languages in parallel corpora\\n will greatly speed up and enhance comparative syntactic research. Automatically\\n extracting and mining syntactic differences from parallel corpora requires a\\n pre-processing step that filters out sentence pairs that cannot be compared\\n syntactically, for example because they involve “free” translations. In this\\n paper we explore four possible filters: the Damerau-Levenshtein distance between\\n POS-tags, the sentence-length ratio, the graph-edit distance between dependency\\n parses, and a combination of the three in a logistic regression model. Results\\n suggest that the dependency-parse filter is the most stable throughout language\\n pairs, while the combination filter achieves the best results.\",\"PeriodicalId\":35138,\"journal\":{\"name\":\"Linguistics in the Netherlands\",\"volume\":\" \",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2019-11-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"1\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Linguistics in the Netherlands\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1075/avt.00029.kro\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"Arts and Humanities\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Linguistics in the Netherlands","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1075/avt.00029.kro","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"Arts and Humanities","Score":null,"Total":0}

引用次数: 1

摘要

并行语料库中大量的语言自动比较将大大加快和加强比较句法研究。从平行语料库中自动提取和挖掘句法差异需要一个预处理步骤，过滤掉无法在句法上进行比较的句子对，例如，因为它们涉及“自由”翻译。在本文中，我们探讨了四种可能的过滤器:pos标签之间的Damerau-Levenshtein距离，句子长度比，依赖解析之间的图形编辑距离，以及三者在逻辑回归模型中的组合。结果表明，依赖解析过滤器在整个语言对中是最稳定的，而组合过滤器的效果最好。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

查看原文本刊更多论文

A filter for syntactically incomparable parallel sentences

Massive automatic comparison of languages in parallel corpora will greatly speed up and enhance comparative syntactic research. Automatically extracting and mining syntactic differences from parallel corpora requires a pre-processing step that filters out sentence pairs that cannot be compared syntactically, for example because they involve “free” translations. In this paper we explore four possible filters: the Damerau-Levenshtein distance between POS-tags, the sentence-length ratio, the graph-edit distance between dependency parses, and a combination of the three in a logistic regression model. Results suggest that the dependency-parse filter is the most stable throughout language pairs, while the combination filter achieves the best results.

求助全文

通过发布文献求助，成功后即可免费获取论文全文。去求助

来源期刊

Linguistics in the Netherlands Arts and Humanities-Language and Linguistics

CiteScore

0.30

自引率

0.00%

发文量

期刊介绍： Linguistics in the Netherlands is a series of annual publications, sponsored by the Dutch Linguistics Association (Algemene Vereniging voor Taalwetenschap) and published by John Benjamins Publishing Company since Volume 8 in 1991. Each volume contains a careful selection through peer review of papers presented at the annual meeting of the society. The aim of the annual meeting is to provide members with an opportunity to report on their work in progress. Each volume presents an overview of research in different fields of linguistics in the Netherlands containing articles on phonetics, phonology, morphology, syntax and semantics.